Hey guys I studied AI in grad school and I was wondering to experiment with coh2 data using some machine learning models. I am specifically interested in supervised learning. Supervised learning is about predicting a value (e.g. Y) from a given set of values (e.g. x1,x2,x3). For example, images can be used to predict or classify a certain output. Given a screenshot (pixel values X) and we could mark where the human player would click (pixel value Y). By learning clicks from good examples, one could train a neural network to predict where should the player(human/AI) click.
This would need a good number of screen shots and collaboration. Just curious to explore this
thanks,
Dedo
Coh2 Research
5 Apr 2021, 21:31 PM
#1
Posts: 22
PAGES (1)
1 user is browsing this thread:
1 guest
Livestreams
78 | |||||
44 | |||||
24 | |||||
17 | |||||
9 | |||||
9 | |||||
0 | |||||
632 | |||||
31 | |||||
17 |
Ladders Top 10
-
#Steam AliasWL%Streak
- 1.653231.739+13
- 2.838223.790+1
- 3.35057.860+15
- 4.590233.717+6
- 5.278108.720+29
- 6.306114.729+2
- 7.645.928+5
- 8.922406.694+1
- 9.1118621.643-1
- 10.265138.658+2
Replay highlight
VS
- cblanco ★
- 보드카 중대
- VonManteuffel
- Heartless Jäger
Einhoven Country
Honor it
10
Download
1243